Functions — github.com/OpenBMB/VisRAG

hub / github.com/OpenBMB/VisRAG / functions

Functions5,024 in github.com/OpenBMB/VisRAG

⨍ Functions 5,024◇ Types & classes 888↳ Endpoints 16

↓ 2 callersFunction_split_args_kwargs_data_proto

(chunks: int, *args, **kwargs)

src/rsgrpo/verl/single_controller/base/decorator.py:48

↓ 2 callersFunction_split_channels

(num_chan, num_groups)

timm_modified/timm/layers/mixed_conv2d.py:14

↓ 2 callersMethod_tokenize

(self, example: str)

src/openmatch/dataset/inference_dataset.py:192

↓ 2 callersFunction_trunc_normal_

(tensor, mean, std, a, b)

timm_modified/timm/layers/weight_init.py:8

↓ 2 callersFunction_try

(syn)

timm_modified/timm/data/dataset_factory.py:50

↓ 2 callersFunction_try_run

(args, initial_batch_size)

timm_modified/validate.py:390

↓ 2 callersFunction_unpad_tensor

(x: Tensor, dim: int, padding_size: int)

src/rsgrpo/verl/utils/ulysses.py:112

↓ 2 callersFunction_unwrap_ray_remote

(cls)

src/rsgrpo/verl/single_controller/ray/base.py:447

↓ 2 callersMethod_update

(self, model, update_fn)

timm_modified/timm/utils/model_ema.py:116

↓ 2 callersFunction_valid_extension

(x: str)

timm_modified/timm/data/readers/img_extensions.py:18

↓ 2 callersFunctionaccuracy_reward

(response: str, ground_truth: str)

src/rsgrpo/examples/reward_function/evidencecot.py:113

↓ 2 callersFunctionadapt_to_chs

(x, n)

timm_modified/timm/data/loader.py:64

↓ 2 callersFunctionadaptive_catavgmax_pool2d

(x, output_size: _int_tuple_2_t = 1)

timm_modified/timm/layers/adaptive_avgmax_pool.py:36

↓ 2 callersFunctionall_gather_data_proto

(data: DataProto, size: int, group: ProcessGroup)

src/rsgrpo/verl/protocol.py:735

↓ 2 callersFunctionall_to_all_tensor

( local_input: Tensor, scatter_dim: int, gather_dim: int, group: Optional[dist.ProcessGroup] =

src/rsgrpo/verl/utils/ulysses.py:134

↓ 2 callersFunctionapply_test_time_pool

(model, config, use_test_size=False)

timm_modified/timm/layers/test_time_pool.py:39

↓ 2 callersFunctionbenchmark

(args)

timm_modified/benchmark.py:575

↓ 2 callersFunctionblockify

image to blocks Args: x (Tensor): with shape (B, H, W, C) block_size (int): edge length of a single square block in units of H, W

timm_modified/timm/models/nest.py:148

↓ 2 callersMethodbottleneck_fn

(self, xs)

timm_modified/timm/models/densenet.py:43

↓ 2 callersMethodbuild

( cls, model_args: ModelArguments, model_name_or_path: str = None, data_args:

src/openmatch/modeling/dense_retrieval_model.py:234

↓ 2 callersFunctionbuild_fourier_pos_embed

Args: feat_shape: Feature shape for embedding. bands: Pre-calculated frequency bands. num_bands: Number of frequency ban

timm_modified/timm/layers/pos_embed_sincos.py:80

↓ 2 callersFunctioncal_rel_pos_type

Spatial Relative Positional Embeddings.

timm_modified/timm/models/mvitv2.py:147

↓ 2 callersMethodcell_forward

(self, x_left, x_right)

timm_modified/timm/models/pnasnet.py:104

↓ 2 callersMethodchannels

feature channels accessor

timm_modified/timm/models/_features.py:69

↓ 2 callersFunctioncheck_batch_size_retry

check failure error string for conditions where batch decay retry should not be attempted

timm_modified/timm/utils/decay_batch.py:32

↓ 2 callersFunctionclean_state_dict

(state_dict: Dict[str, Any])

timm_modified/timm/models/_helpers.py:22

↓ 2 callersFunctioncompute_kl

Compute KL divergence given log_probs and ref_log_probs. Adapted from https://github.com/huggingface/trl/blob/v0.11.0/trl/trainer/ppo_trainer.py#

src/rsgrpo/verl/trainer/core_algos.py:523

↓ 2 callersFunctioncompute_length_metrics

(batch: DataProto)

src/rsgrpo/verl/trainer/metrics.py:27

↓ 2 callersMethodcompute_log_prob

Compute logits given a batch of data. Args: data (DataProto): a batch of data represented by DataProto. It must contain key ```in

src/rsgrpo/verl/workers/actor/base.py:40

↓ 2 callersFunctionconv2d_same

( x, weight: torch.Tensor, bias: Optional[torch.Tensor] = None, stride: Tuple[

timm_modified/timm/layers/conv2d_same.py:17

↓ 2 callersFunctionconvert

(mxnet_name, torch_name)

timm_modified/convert/convert_from_mxnet.py:15

↓ 2 callersFunctioncreate_conv2d_pad

(in_chs, out_chs, kernel_size, **kwargs)

timm_modified/timm/layers/conv2d_same.py:96

↓ 2 callersFunctioncreate_reader

( name: str, root: Optional[str] = None, split: str = 'train', **kwargs, )

timm_modified/timm/data/readers/reader_factory.py:8

↓ 2 callersFunctioncreate_resnetv2_stem

( in_chs, out_chs=64, stem_type='', preact=True, conv_layer=StdConv2d,

timm_modified/timm/models/resnetv2.py:279

↓ 2 callersFunctioncreate_scheduler_v2

( optimizer: Optimizer, sched: str = 'cosine', num_epochs: int = 300, decay_ep

timm_modified/timm/scheduler/scheduler_factory.py:63

↓ 2 callersFunctioncreate_shortcut

( downsample_type, in_chs, out_chs, kernel_size, stride, dilat

timm_modified/timm/models/regnet.py:151

↓ 2 callersFunctiondeblockify

blocks to image Args: x (Tensor): with shape (B, T, N, C) where T is number of blocks and N is sequence size per block block_size

timm_modified/timm/models/nest.py:165

↓ 2 callersFunctiondecay_batch_step

power of two batch-size decay with intra steps Decay by stepping between powers of 2: * determine power-of-2 floor of current batch size (ba

timm_modified/timm/utils/decay_batch.py:8

↓ 2 callersFunctiondispatch_clip_grad

Dispatch to gradient clipping method Args: parameters (Iterable): model parameters to clip value (float): clipping value/factor/

timm_modified/timm/utils/clip_grad.py:6

↓ 2 callersFunctiondistributed_parallel_embedding_inference

( dataset: InferenceDataset, model: DRModelForInference, args: EncodingArguments, dataset_type

src/openmatch/inference/inference.py:53

↓ 2 callersFunctiondownload_cached_file

(url, check_hash=True, progress=False)

timm_modified/timm/models/_hub.py:71

↓ 2 callersFunctionencode

(text_or_image_list)

visrag_scripts/demo/retriever/demo.py:15

↓ 2 callersFunctionencode

(model, tokenizer, text_or_image_list)

visrag_scripts/demo/visrag_pipeline/utils.py:13

↓ 2 callersMethodestimate_flops

Estimate the FLOPS based on the number of valid tokens in the current batch and the time taken. Args: batch_seqlens (Lis

src/rsgrpo/verl/utils/flops_counter.py:118

↓ 2 callersMethodexecute_all_async

(self, method_name: str, *args, **kwargs)

src/rsgrpo/verl/single_controller/ray/base.py:372

↓ 2 callersMethodexecute_rank_zero_async

(self, method_name: str, *args, **kwargs)

src/rsgrpo/verl/single_controller/ray/base.py:359

↓ 2 callersFunctionf1_match

(response: str, ground_truth: str)

src/rsgrpo/examples/reward_function/evidencecot.py:89

↓ 2 callersFunctionfilter_pretrained_cfg

(cfg, remove_source=False, remove_null=True)

timm_modified/timm/models/_pretrained.py:69

↓ 2 callersFunctionfind_all_markers

Find all markers' names (quoted in "<>") in a template.

src/openmatch/utils.py:208

↓ 2 callersFunctionflatten_dict

(data: Dict[str, Any], parent_key: str = "", sep: str = "/")

src/rsgrpo/verl/utils/py_functional.py:106

↓ 2 callersFunctionflatten_modules

( named_modules: Iterator[Tuple[str, nn.Module]], depth: int = 1, prefix: Union[str, T

timm_modified/timm/models/_manipulate.py:156

↓ 2 callersFunctionformat_reward

(response: str, ground_truth: str)

src/rsgrpo/examples/reward_function/evidencecot.py:160

↓ 2 callersMethodforward_cls

(self, x)

timm_modified/timm/models/volo.py:631

↓ 2 callersMethodforward_features

(self, x)

timm_modified/timm/models/inception_v3.py:365

↓ 2 callersMethodforward_features

(self, x0)

timm_modified/timm/models/coat.py:570

↓ 2 callersMethodforward_head

(self, x)

timm_modified/timm/models/inception_v3.py:374

↓ 2 callersMethodforward_postaux

(self, x)

timm_modified/timm/models/inception_v3.py:359

↓ 2 callersMethodforward_tokens

(self, x)

timm_modified/timm/models/volo.py:616

↓ 2 callersFunctiongather_outputs_and_unpad

Gather a tensor across a process group and optionally unpad its padded elements. Args: x (Tensor): Input tensor to gather. g

src/rsgrpo/verl/utils/ulysses.py:238

↓ 2 callersMethodgather_tensor

(self, t)

src/openmatch/loss.py:47

↓ 2 callersFunctiongen_relative_position_index

( q_size: Tuple[int, int], k_size: Optional[Tuple[int, int]] = None, class_token: bool

timm_modified/timm/layers/pos_embed_rel.py:21

↓ 2 callersFunctiongen_relative_position_index

(window_size: Tuple[int, int])

timm_modified/timm/models/beit.py:61

↓ 2 callersFunctiongenerate_lookup_tensor

Generate a one_hot lookup tensor to reindex embeddings along one dimension. Args: length: the length to reindex to. max_relative_

timm_modified/timm/layers/pos_embed_rel.py:399

↓ 2 callersFunctiongetPublicAttr

(obj)

timm_modified/tests/test_optim.py:133

↓ 2 callersFunctionget_1d_sincos_pos_embed_from_grid

embed_dim: output dimension for each position pos: a list of positions to be encoded: size (M,) out: (M, D)

src/openmatch/modeling/weighted_selection/MiniCPMV20/resampler.py:87

↓ 2 callersFunctionget_1d_sincos_pos_embed_from_grid

embed_dim: output dimension for each position pos: a list of positions to be encoded: size (M,) out: (M, D)

src/openmatch/modeling/modeling_minicpmv/resampler.py:72

↓ 2 callersFunctionget_2d_sincos_pos_embed

grid_size: int of the grid height and width return: pos_embed: [grid_size*grid_size, embed_dim] or [1+grid_size*grid_size, embed_dim] (w/

src/openmatch/modeling/weighted_selection/MiniCPMV20/resampler.py:53

↓ 2 callersFunctionget_2d_sincos_pos_embed

grid_size: int of the grid height and width return: pos_embed: [grid_size*grid_size, embed_dim] or [1+grid_size*grid_size, embed_dim] (w/

src/openmatch/modeling/modeling_minicpmv/resampler.py:38

↓ 2 callersMethodget_accuracy

(self, k=None)

timm_modified/timm/data/real_labels.py:43

↓ 2 callersMethodget_attention

(self, x)

timm_modified/timm/models/convit.py:80

↓ 2 callersMethodget_attention_biases

(self, device: torch.device)

timm_modified/timm/models/levit.py:209

↓ 2 callersMethodget_attention_biases

(self, device: torch.device)

timm_modified/timm/models/levit.py:310

↓ 2 callersFunctionget_cache_dir

Returns the location of the directory where models are cached (and creates it if necessary).

timm_modified/timm/models/_hub.py:56

↓ 2 callersMethodget_num_gpus

Get the number of gpus in this cluster.

src/rsgrpo/verl/trainer/ray_trainer.py:98

↓ 2 callersMethodget_passage

(self, p)

src/openmatch/utils.py:79

↓ 2 callersMethodget_process_fn

(self, epoch, hashed_seed)

src/openmatch/dataset/train_dataset.py:61

↓ 2 callersFunctionget_rel_pos

Get relative positional embeddings according to the relative positions of query and key sizes. Args: q_size (int): size of qu

timm_modified/timm/models/vision_transformer_sam.py:37

↓ 2 callersFunctionget_seqlen_balanced_partitions

Get order of seq lengths to make partitions balanced, this is used in balacing sum of seqlength across dp ranks and microbatches. Parameters:

src/rsgrpo/verl/utils/seqlen_balancing.py:153

↓ 2 callersMethodget_slice_image_placeholder

(self, image, tokenizer)

src/openmatch/modeling/modeling_minicpmv/modeling_minicpmv.py:247

↓ 2 callersMethodget_vision_embedding

(self, pixel_values)

src/openmatch/modeling/weighted_selection/MiniCPMV20/modeling_minicpmv.py:86

↓ 2 callersMethodget_vllm_embedding

(self, data)

src/openmatch/modeling/weighted_selection/MiniCPMV20/modeling_minicpmv.py:103

↓ 2 callersMethodget_vllm_embedding

(self, data)

src/openmatch/modeling/modeling_minicpmv/modeling_minicpmv.py:124

↓ 2 callersFunctiongrid_partition

(x, grid_size: List[int])

timm_modified/timm/models/maxxvit.py:650

↓ 2 callersFunctiongrid_reverse

(windows, grid_size: List[int], img_size: List[int])

timm_modified/timm/models/maxxvit.py:660

↓ 2 callersFunctiongroup_rms

(x, groups: int = 32, eps: float = 1e-5)

timm_modified/timm/layers/evo_norm.py:90

↓ 2 callersFunctiongroup_with_matcher

( named_objects: Iterator[Tuple[str, Any]], group_matcher: Union[Dict, Callable], retu

timm_modified/timm/models/_manipulate.py:75

↓ 2 callersFunctionhf_split

(hf_id: str)

timm_modified/timm/models/_hub.py:116

↓ 2 callersMethodindex_to_description

(self, index: int, detailed: bool = False)

timm_modified/timm/data/imagenet_info.py:86

↓ 2 callersMethodindex_to_label_name

(self, index)

timm_modified/timm/data/imagenet_info.py:81

↓ 2 callersMethodinit_weights

(self, scheme='')

timm_modified/timm/models/maxxvit.py:1093

↓ 2 callersMethodinit_weights

(self)

timm_modified/timm/models/swin_transformer_v2_cr.py:317

↓ 2 callersMethodinit_weights

(self, nlhb=False)

timm_modified/timm/models/sequencer.py:396

↓ 2 callersMethodinit_weights

(self, nlhb=False)

timm_modified/timm/models/mlp_mixer.py:234

↓ 2 callersFunctioninplace_abn

(x, weight, bias, running_mean, running_var, training=True, momentum=0.1, eps=1e-05, activ

timm_modified/timm/layers/inplace_abn.py:10

↓ 2 callersFunctioninstance_rms

(x, eps: float = 1e-5)

timm_modified/timm/layers/evo_norm.py:47

↓ 2 callersFunctioninv_instance_rms

(x, eps: float = 1e-5)

timm_modified/timm/layers/filter_response_norm.py:14

↓ 2 callersFunctionisObserve_punishment

(response: str)

src/rsgrpo/examples/reward_function/evidencecot.py:201

↓ 2 callersFunctionisThink_punishment

(response: str)

src/rsgrpo/examples/reward_function/evidencecot.py:210

↓ 2 callersFunctionis_rank0

()

src/rsgrpo/verl/utils/model_utils.py:27

↓ 2 callersFunctionis_stem_deep

(stem_type)

timm_modified/timm/models/resnetv2.py:275

← previous next →601–700 of 5,024, ranked by callers